Meet the New Pirates
On Napster, Spotify, and the corporate acquisition of someone else's music
The story the music industry tells about the streaming era has a clean narrative shape. There was a crisis. Napster, peer-to-peer file sharing, the decade in which people simply took music without paying for it and the industry watched its revenue collapse. Then the resolution. The cavalry came. Spotify. Legal streaming scaled. The pirates were defeated. The industry, battered but surviving, adapted to a new model and moved on.
It is a satisfying story. It also has a significant problem, which is that the resolution it describes did not occur. The pirates were not defeated. They were replaced by a more sophisticated, more profitable version of the same arrangement — one in which music is still used without meaningful compensation to the people who made it, the infrastructure for taking it is still controlled by people who did not make it, and the main difference is that the entity extracting value from other people's creative labour now has a stock market listing and a terms of service agreement.
Artists and industry critics have been making versions of this argument for as long as streaming has existed. What has changed is the accumulation of evidence, over two decades, that the streaming model was not a transitional arrangement that would eventually generate fair returns as the market matured. It was the final destination. The question worth examining is how we arrived here, and why the story of rescue was so effective for so long.
Napster launched in 1999 and was shut down by court order in 2001, after the Recording Industry Association of America successfully argued that it facilitated mass copyright infringement.(1) The legal argument was straightforward: Napster provided infrastructure that allowed users to share copyrighted music files without authorisation or compensation to rights holders. The fact that Napster itself did not host the files, and that the sharing occurred between users, with Napster providing only the indexing and search infrastructure, was insufficient to insulate it from liability.
The cultural argument against Napster was similarly direct, and it was made with considerable force by artists who understood precisely what was happening to their work. Their music was being copied and distributed without their consent and without payment. The people doing it had rationalised this in various ways — information wants to be free, the labels were the real thieves, music should be available to everyone regardless of ability to pay — but the rationalisation did not change the underlying transaction, which was the use of someone else's creative labour without compensation.
Daniel Ek has described Spotify's founding thesis in multiple interviews: streaming, offered at a price point low enough to compete with free, would return music listeners to legal consumption and restore revenue to the industry.(2) This argument was persuasive and, in narrow terms, accurate. Spotify did reduce piracy in the markets where it launched. Listeners who had been taking music without paying began paying — a small monthly fee, distributed across everything they streamed, returning fractions of fractions to the rights holders whose catalogues they accessed.
The rights holders who benefited most from this arrangement were the major labels, who negotiated licensing deals with Spotify that included equity stakes in the platform — meaning they participated not only in the royalty stream but in the capital appreciation of the company they were licensing their catalogues to.
When Spotify went public in 2018, Universal, Sony, and Warner collectively held stakes worth hundreds of millions of dollars, acquired in exchange for the catalogue access that made the platform viable in its early years. The labels had been pirates once too. Now they were shareholders.
Independent artists and smaller rights holders participated in a different version of this arrangement. They received royalties calculated under the pro-rata model — a share of the total royalty pool proportional to their share of total streams — without equity, without the negotiating leverage to influence the royalty rate, and without the ability to opt out of the platform without accepting the commercial consequences of absence from the dominant distribution channel. The rescue had terms. The terms were not available to everyone equally.
Piracy, as a legal and ethical concept, describes the use of someone's creative work without their consent and without fair compensation. The emphasis on both elements matters: consent without compensation is exploitation, and compensation without consent is still a violation of the creator's right to determine how their work is used. The moral wrong of piracy is not only that creators go unpaid. It is that the decision about whether and how to distribute their work is made by someone else. Measured against this definition, the streaming model deserves more scrutiny than it typically receives. The royalty rates paid to independent artists by major streaming platforms represent compensation so far below the value generated that the word compensation begins to strain.
A system that takes someone's creative work, distributes it to hundreds of millions of listeners, generates billions of dollars in platform revenue and shareholder returns, and delivers the creator less than $300 a year in income is not obviously distinguishable from a system that takes someone's creative work without paying them.
Artists who wish to reach listeners in the streaming era have no practical alternative to the major platforms. Spotify, Apple Music, and Amazon Music collectively account for the substantial majority of global streaming consumption.(3) An artist who declines to participate on the available terms has the theoretical freedom to make that choice and the practical consequence of disappearing from the channels through which most listeners discover and consume music.
Liz Pelly's reporting in Mood Machine documented Spotify's use of its own recommendation infrastructure to promote music it had commissioned from production partners at below-standard royalties, flooding the functional music categories with generic AI-assisted productions that occupied algorithmic recommendation slots at the expense of independent artists working in the same space.(4) The platform was not merely distributing other people's music at below-fair rates. It was using its control of the distribution infrastructure to compete with the artists whose presence on the platform gave that infrastructure its value.
The AI training data question extends this logic further. The major AI music generation systems now available commercially were trained on recorded music — the accumulated creative output of human musicians across decades — without the consent of those musicians and without compensation.(5) The legal status of this appropriation is contested, with multiple lawsuits in progress in various jurisdictions. The ethical status is less contested: the creative labour of specific human beings was used, without their knowledge or permission, to train systems that now compete with them in the market for recorded music.
The response to this argument tends to be more legally nuanced and philosophically elaborate than the response Napster's defenders offered, more attentive to the distinction between inspiration and reproduction. But the underlying transaction is recognisable. Someone else's creative work was used to build something profitable, without consent and without compensation.
The rescue narrative was convincing for as long as it was because it contained a genuine element of truth. Streaming did reduce piracy. It did return some revenue to some rights holders. It did make music more accessible to more listeners at a lower cost than the preceding model.
What the narrative did not survive was the accumulation of evidence about who, specifically, was rescued and on what terms. The major labels were rescued of course, with equity stakes, favourable licensing terms, and a distribution model that concentrated royalty income at the top of the catalogue. The artists whose work those catalogues contained weren’t rescued they were held to ransom, with no practical recourse when the terms of the ransom proved unfavourable.
The complaint against Napster was not that music was unavailable or that listeners were not enjoying it. Music was very available. Listeners were enjoying it enormously. The complaint was that the people who made it were not being compensated fairly for its use, and that the infrastructure enabling that use was controlled by people with no stake in the creative work's existence.(6) Twenty-five years later, music is still very available. Listeners are still enjoying it enormously. The people who make it are still not being compensated fairly for its use. The infrastructure enabling that use is still controlled by people whose primary stake is financial rather than creative.
The distinction between the Napster model and the Spotify model that actually matters is not the presence of payment. It is the question of control; who decides how the music is used, how it is presented, how it is recommended, what it is worth, and what happens to the data generated by its consumption.
What genuine rescue should have looked like — what it could still look like — is a distribution infrastructure in which artists have not merely legal ownership of their work but practical rights over how it is used. Where the people who make the music have a meaningful voice in the decisions that determine what it is worth, how it is presented, and what the platform does with the returns. Where the question of whether the arrangement is fair is answered by the people most affected by it rather than by the shareholders of the entity extracting value from it.
This is the cooperative argument, and it’s not a utopian alternative to commercial reality. If the great evil of piracy is that someone else's creative work is used without their consent and without fair compensation, then the solution is not merely about payment, it is about creating the structural conditions under which consent is meaningful and compensation is determined by the people whose work is being valued. Spotify was never the cavalry. Platform capitalists were never going to do this work.
The streaming era has taught the music industry that the problem was never really about the technology. The problem was always who controlled the infrastructure through which music reached listeners, and whose interests that infrastructure was designed to serve. And since the first labels emerged those interests have been the same. The pirates wore eye patches. Their successors wear Armani suits. The music is still being taken.
Notes
RIAA v. Napster, 239 F.3d 1004 (9th Cir. 2001). https://law.justia.com/cases/federal/appellate-courts/F3/239/1004/600778/
Daniel Ek on Spotify's founding thesis — Rolling Stone (2012) and various investor presentations. https://www.rollingstone.com
Midia Research and IFPI Global Music Report 2024 — streaming market share data. https://www.ifpi.org/resources/
Liz Pelly, Mood Machine (Atria Books, 2025). https://www.simonandschuster.com/books/Mood-Machine/Liz-Pelly/9781668083949
AI music training data litigation — ongoing in multiple jurisdictions. https://www.musicbusinessworldwide.com
The RIAA's anti-piracy campaigns positioned labels as defenders of artist welfare. On the subsequent streaming deal terms, see Robert Levine, Free Ride (Anchor Books, 2012).